Computer Vision/Machine Learning Engineer

Summary

Join Canvas, a leading-edge AI and computer vision technology company revolutionizing the remodeling, architecture, and interior design industry. We use LiDAR-enabled scans to create precise 3D models of homes. As a Computer Vision / Machine Learning Engineer, you will significantly impact our business by automating our Scan-to-CAD process using our extensive dataset. This role blends research and engineering, from reviewing papers and experimenting to deploying solutions and integrating them into our product. You will collaborate with various teams to improve efficiency and contribute to our core innovation team. We are a global virtual-first company with a distributed team.

Requirements

• Deep expertise and extensive hands-on experience with modern machine learning techniques in the field of computer vision
• Rapid iteration and experimentation skills, ability to quickly evaluate ideas and test their fit to the product needs, ability to plan, setup and implement large-scale experiments
• Strong skills of working with research literature: you should be ready to review dozens of papers within days; you should be able to grasp key concepts and ideas from papers quickly and efficiently
• Good knowledge of Python and PyTorch
• Strong communication skills, fluent English
• Comfortable working across multiple time zones and cultures

Responsibilities

• Review research papers and experiment with state-of-the-art approaches in the field of 3D scene understanding and Scan-to-CAD conversion
• Develop new algorithms and train neural networks to solve the Scan-to-CAD conversion problem both end-to-end and in parts
• Collaborate closely with the 3D Operations, Visualization & Tooling and other teams to improve the efficiency of manual Scan-to-CAD conversion by automating our internal 3D tooling and integrating developed CV/ML solutions into the production pipeline
• Improve the ML infrastructure: Set up efficient data pipelines and automate training and deployment processes

Preferred Qualifications

• MS or PhD with specialization in machine learning or computer vision, or relevant experience in academia
• Publications on 1st-tier computer vision conferences (CVPR, ICCV, ECCV)
• Experience with machine learning in application to the 3D domain, especially the problem of 3D scene understanding, but also SLAM, 3D reconstruction, depth estimation, and similar
• Knowledge of MLOps best practices at least on application level
• Understanding of 3d model representation